The Joint Multivariate Modeling of Multiple Mixed Response Sources: Relating Student Performances with Feedback...

Area of Science:

• Educational Technology
• Psychometrics
• Learning Analytics

Background:

• Computer-based assessments are increasingly used to measure student skills.
• Information literacy is a critical skill for academic success.
• Effective feedback mechanisms are essential for student learning and improvement.

Purpose of the Study:

• To analyze student feedback behavior, including usage and attention, in a Dutch computer-based assessment.
• To investigate the relationship between test performance, working speed, and feedback utilization.
• To develop and apply a multivariate hierarchical latent variable model for analyzing complex student data.

Main Methods:

• Utilized a multivariate hierarchical latent variable model to analyze data from a Dutch computer-based assessment.
• Examined student feedback behavior (use and attention time), test performance, and working speed.
• Employed a flexible within-subject latent variable structure to capture individual differences.

Main Results:

• Well-performing students tended to visit feedback pages less frequently for easier items.
• Students' attention to feedback was positively associated with their working speed.
• Attention paid to feedback did not predict the likelihood of students using the feedback.

Conclusions:

• Student engagement with feedback in computer-based assessments is nuanced and influenced by item difficulty and performance level.
• Working speed is a significant factor in how students interact with feedback, but not necessarily in their decision to utilize it.
• The proposed latent variable model effectively captures complex relationships between performance, behavior, and feedback in educational technology settings.